Test-retest reliability of simulated driving performance: A pilot study

Abstract

Twenty-seven volunteers completed three simulated driving tests to determine test-retest reliability of performance on a low-cost, fixed-base computerized driving simulator. One retest was completed a few hours after the initial drive, and the final retest was completed 7 days following the initial test drive. Driving performance was compared using measures of vehicle control, speed, and reaction time to critical events. A measure of participants' ability to inhibit a pre-potent response was also assessed using an inhibition task during each drive, with the number of incorrect inhibition responses recorded. Practice effects were evident for measures of vehicle control (deviation of lane position and number of line crossings) and participants' ability to withhold responses to inhibition tasks. Good test-retest reliability was observed for measures of vehicle control, speed, reaction time, and variability measures. Poor test-retest reliability was observed for the number of stopping failures observed during driving. The findings from this study suggest that the driving scenario used provides reliable assessment tasks that could be used to track the effects of pharmacological treatments on driving performance. However, an additional familiarization drive should be included as part of future study protocols employing this driving scenario to reduce learning effects during trials. Care should also be taken when interpreting results from tasks with low test-retest reliability.